maxwell-feng/dsh-tesseract-ocr ↗★ 2
@maxwell-feng/dsh-tesseract-ocr
dsh plugin: recognize attached images locally with Tesseract OCR and send only the recognized text to the model. Text models never receive image bytes; vision passthrough is opt-in.
AI Analysis
核心用途是让不支持多模态的纯文本模型也能“看懂”图片中的文字。适合希望节省 Vision 模型 Token 消耗，或在本地离线处理图片文字识别任务的用户。
Install
$
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:maxwell-feng/dsh-tesseract-ocrREADME
Read the full README ↗Configuration
All settings live in the patch row tesseract-ocr:
| Key | Default | Meaning |
|---|---|---|
language | eng | Tesseract language(s), +-joined, e.g. eng, chi_sim, eng+chi_sim |
passthrough | false | false (default): OCR every image. true: genuine vision models receive images untouched |
tesseractBin | tesseract | CLI path; quote paths with spaces, e.g. "C:\Program Files\Tesseract-OCR\tesseract.exe" |
psm | 3 | Page segmentation mode (tesseract --psm) |
timeoutMs | 60000 | Per-image OCR timeout |
maxCacheEntries | 200 | Bound on the per-run OCR cache (keyed by attachment id) |